CN112598486A - Marketing accurate screening push system based on big data and intelligent Internet of things - Google Patents

Marketing accurate screening push system based on big data and intelligent Internet of things Download PDF

Info

Publication number
CN112598486A
CN112598486A CN202110019917.6A CN202110019917A CN112598486A CN 112598486 A CN112598486 A CN 112598486A CN 202110019917 A CN202110019917 A CN 202110019917A CN 112598486 A CN112598486 A CN 112598486A
Authority
CN
China
Prior art keywords
customer
module
heat
information
data
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202110019917.6A
Other languages
Chinese (zh)
Other versions
CN112598486B (en
Inventor
张胜敏
刘悦
张书贵
徐慧玲
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Kaifeng University
Original Assignee
Kaifeng University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Kaifeng University filed Critical Kaifeng University
Priority to CN202110019917.6A priority Critical patent/CN112598486B/en
Publication of CN112598486A publication Critical patent/CN112598486A/en
Application granted granted Critical
Publication of CN112598486B publication Critical patent/CN112598486B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/06—Buying, selling or leasing transactions
    • G06Q30/0601—Electronic shopping [e-shopping]
    • G06Q30/0631—Item recommendations
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/953—Querying, e.g. by the use of web search engines
    • G06F16/9535—Search customisation based on user profiles and personalisation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/953—Querying, e.g. by the use of web search engines
    • G06F16/9537—Spatial or temporal dependent retrieval, e.g. spatiotemporal queries
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V2201/00—Indexing scheme relating to image or video recognition or understanding
    • G06V2201/07—Target detection
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Business, Economics & Management (AREA)
  • Computing Systems (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Evolutionary Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Health & Medical Sciences (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Software Systems (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Finance (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Accounting & Taxation (AREA)
  • Mathematical Physics (AREA)
  • Development Economics (AREA)
  • General Business, Economics & Management (AREA)
  • Economics (AREA)
  • Marketing (AREA)
  • Strategic Management (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention relates to the technical field of intelligent marketing, big data and intelligent Internet of things, in particular to a marketing accurate screening and pushing system based on big data and intelligent Internet of things. The system comprises: the system comprises a market perception module, a market operation big data module, a data calculation analysis module and a commodity pushing module; obtaining customer information and sales data in a market through a market perception module and a market operation big data module; the data calculation and analysis module processes the customer information and sales data to obtain a shop area characteristic matrix; and the commodity pushing module analyzes the customer information and the shop area characteristic matrix and pushes commodities to the customer. According to the invention, commodity pushing is carried out on the customer in an entity economic system under the online by the intelligent Internet of things and big data analysis, so that the consumption capacity of the customer is promoted.

Description

Marketing accurate screening push system based on big data and intelligent Internet of things
Technical Field
The invention relates to the technical field of intelligent marketing, big data and intelligent Internet of things, in particular to a marketing accurate screening and pushing system based on big data and intelligent Internet of things.
Background
With the development of scientific technology, new products and new services of the big data market are continuously emerging, and big data becomes a current hot topic. However, the rapid development of the online economy causes the delay of the development of the offline physical economy and the reduction of the operation, but the physical stores are also indispensable in daily life, so that the joint development of the physical economy is also emphasized while the online economy is developed. Meanwhile, the online shopping has many defects, such as inconsistent pictures with real objects, unqualified quality and the like. How to promote customer consumption capacity through big data analysis in an entity economic system under line is a problem to be solved at present.
Disclosure of Invention
In order to solve the technical problems, the invention aims to provide a marketing accurate screening and pushing system based on big data and an intelligent internet of things, and the technical scheme is as follows:
the invention provides a marketing accurate screening and pushing system based on big data and an intelligent Internet of things, which comprises: the system comprises a market perception module, a market operation big data module, a data calculation analysis module and a commodity pushing module;
the shopping mall sensing module is used for acquiring customer information through monitoring equipment in a shopping mall; the customer information includes customer location information;
the mall operation big data module is used for acquiring operation data of the mall, and the operation data comprises sales volume data of shops inside the mall;
the data calculation and analysis module is used for projecting and transforming the customer position information to a preset shopping mall plane overlook image and obtaining a projection heat map of the shopping mall area through a projection point; discrete data points representing heat are included in the projected heat map; building a Thiessen polygon through the discrete data points; calculating the peak heat value of each peak in the Tassen polygon through an interpolation algorithm; obtaining the influence degree of the vertex on the heat of any position by calculating the distance of the vertex heat value to any position in the heat map so as to obtain the heat of any position; obtaining a heat grade characteristic diagram of the shopping mall according to the heat of the any position; converting the sales data into a sales proportion matrix; combining the sales rate proportion matrix and the heat level characteristic map into a shop area characteristic matrix;
the commodity pushing module is used for pushing commodities to the customer through display equipment in combination with the customer information and the shop area characteristic matrix analysis.
Further, the market perception module further comprises a customer behavior detection module;
the customer behavior detection module is used for judging whether the customer enters a fitting room or not through a pedestrian re-identification technology and recording the times of the customer entering the fitting room; identifying the label brought into the clothing hangtag of the fitting room by using a wireless video technology to obtain a fitting list taken by the customer; the fitting list comprises the identified number of clothes and the clothes attribute; the times of the customer entering the fitting room and the fitting list are used as customer behavior information; the customer information includes the customer behavior information.
Furthermore, the market perception module further comprises a customer enclosure acquisition module and an enclosure analysis module;
the customer enclosure frame acquisition module is used for outputting a target enclosure frame of the customer through a pre-trained target detection network;
the bounding box analysis module is used for analyzing the target bounding box through a pre-trained convolutional neural network to obtain the physique information of the customer; the customer information comprises the customer physique information; obtaining customer clothing matching information through a pre-trained example segmentation network; the customer information comprises the customer clothing matching information.
Further, the data calculation and analysis module further comprises a heat map acquisition module;
the heat map acquisition module is used for carrying out four-point method estimation on the ground labeling of the image acquired by the monitoring equipment and the corresponding coordinates of the top plan image of the shopping mall, projecting the bottom edge center of the target enclosure frame into the top plan image of the shopping mall through a homography matrix and acquiring the projection point; generating a thermodynamic diagram with two-dimensional Gaussian distribution based on the projection points in the planar top-view image of the mall; and obtaining the projection heat map of the market area according to time sequence statistics.
Further, the data calculation and analysis module further comprises an information filtering module;
the information filtering module is used for filtering the heat map by a maximum lattice point sampling method; and processing the projection heat map through a preset sliding window, wherein only the maximum value in the sliding window is reserved in each processing, and other values are subjected to zero returning processing.
Further, the data calculation and analysis module further comprises a region heat acquisition module;
the region heat acquisition module is used for carrying out classification region division on the top plan view of the market to obtain a region image; calculating the area of any region in the region image, and calculating the region heat through the number, the area and discrete data point values of the Thiessen polygons contained in the region:
Figure BDA0002888272170000021
wherein, AreaHiRepresenting the heat, S, of the ith area in the area imagei,jRepresenting the area of the jth Thiessen polygon in the ith area, SiRepresenting the area of the ith said region in said region image, Hi,jDiscrete data point values representing the jth Thiessen polygon contained in the ith region, n representing n regions divided in the region image.
Further, the data calculation and analysis module further comprises a cluster analysis module;
the cluster analysis module is used for generating a market heat map after obtaining the heat of any position, and dividing the heat level of the market heat map based on pixel values through a clustering algorithm to generate the heat level characteristic map.
Further, the commodity pushing module further comprises a commodity pushing neural network module;
and the commodity pushing neural network module is used for analyzing the customer information and the shop area characteristic matrix input into the network through the trained commodity pushing neural network and outputting the pushed commodities.
Further, the commodity pushing module further comprises an offline commodity pushing module;
the off-line commodity pushing module is used for counting the heat and the stay time of the area in the market where the track information is located by analyzing the track information of the customer in the market, and pushing the commodities through a mobile phone after the customer leaves the market.
The invention has the following beneficial effects:
1. according to the embodiment of the invention, the information of the customer and the information in the market are constructed through the big data of the visual perception in the market and the perception of the intelligent Internet of things. Through the analysis and training of the information, the system can analyze the current information in real time and accurately push commodities to customers.
2. The commodity pushing neural network in the embodiment of the invention only needs to acquire the labeling data based on the historical sales data, the visual perception data and the perception data of the Internet of things, the training method is simple, the data is easy to acquire, and the label data does not need to be marked artificially.
3. According to the embodiment of the invention, the movement track information of the customer in the market is analyzed through the off-line pushing module, and the off-line commodity pushing can be carried out on the customer according to the track information when the customer leaves the market and needs to push commodities, so that the pushing efficiency is improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions and advantages of the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
Fig. 1 is a block diagram of a marketing precision screening and pushing system based on big data and an intelligent internet of things according to an embodiment of the present invention;
fig. 2 is a block diagram of a commodity push neural network according to an embodiment of the present invention.
Detailed Description
In order to further illustrate the technical means and effects of the present invention adopted to achieve the predetermined purpose, the following detailed description, with reference to the accompanying drawings and preferred embodiments, describes specific embodiments, structures, features and effects of a marketing precision screening and pushing system based on big data and intelligent internet of things according to the present invention. In the following description, different "one embodiment" or "another embodiment" refers to not necessarily the same embodiment. Furthermore, the particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The specific scheme of the marketing accurate screening and pushing system based on big data and an intelligent internet of things is specifically described below with reference to the accompanying drawings.
Referring to fig. 1, a block diagram of a marketing precision screening and pushing system based on big data and an intelligent internet of things according to an embodiment of the present invention is shown, where the system includes: the system comprises a market perception module 101, a market operation big data module 102, a data calculation and analysis module 103 and a commodity pushing module 104.
The mall sensing module 101 is used for acquiring customer information through monitoring equipment inside a mall. The monitoring equipment comprises a monitoring camera, a sensor and other equipment. The customer information comprises customer physique information, customer behavior information and customer clothing matching information.
In the embodiment of the invention, the physical information of the customer comprises the sex, the age, the height and the weight grade of the customer; the customer behavior information comprises the times of entering the fitting room and fitting lists of the customer; the customer clothing matching information comprises the color, style, fabric and style of the clothing worn by the customer.
Preferably, the mall sensing module 101 in the embodiment of the present invention includes a customer enclosure acquisition module and an enclosure analysis module.
The customer enclosure frame acquisition module is used for outputting the target enclosure frame of the customer through a pre-trained target detection network. The training process of the target detection network specifically comprises the following steps:
1) and using the images collected by the monitoring equipment in the market as training images. And marking the position of the center point of the customer in the training image and the length and width information of the surrounding frame. And (4) convolving the customer center position through a Gaussian convolution core to obtain a customer center heat map. There is (x, y, w, h) information for the location of each point in the customer-centric heatmap, x, y being the coordinates of the point representing the customer, and w, h being the length and width of the bounding box.
2) And (4) normalizing the training image and the label data to change the image matrix into a floating point number between [0 and 1] so as to facilitate better convergence of the model. And inputting the normalized data into a network.
3) The network structure employs an encoding-decoding structure. And the target detection encoder performs feature extraction on the training image and outputs a feature map. And the target detection decoder performs up-sampling and feature extraction on the feature map and outputs the customer center heat map and the bounding box.
4) The network uses a weighted sum of the central point prediction penalty and the bounding box size penalty. The mathematical formula for the center point loss is as follows:
Figure BDA0002888272170000051
wherein alpha and beta are holothuriansN is the number of center points of the bounding box in the image, set by human experience. Gamma rayxyFor the predicted value of xy coordinates in the customer-centric heatmap, yxyIs the value of xy coordinates in the real data Heatmap (group Truth Heatmap).
The mathematical formula for the bounding box size penalty is as follows:
Figure BDA0002888272170000052
wherein N is the number of center points of the surrounding frame in the image, SPkFor the predicted length and width of the bounding box, skThe length and width of the box are enclosed for real data.
The overall loss function is:
Total Loss＝CenterLoss+δ*SizeLoss
where δ is the weight. In the present example, δ is 0.1.
5) And carrying out post-processing on the obtained customer center heat map and the surrounding frame, searching a peak point, and obtaining specific target surrounding frame information and center point coordinate information as customer position information. Post-processing methods include maximum suppression, Softargmax, and the like.
The bounding box analysis module is used for analyzing the target bounding box through a pre-trained convolutional neural network to obtain the physique information of the customer, and the convolutional neural network can be analyzed by adopting a residual error network (ResNet). And analyzing the images in the target enclosure frame through a pre-trained example segmentation network to obtain the segmented images of each piece of clothing on the customer. Multiplying the segmented image with the original image to obtain the original image of the clothing example without background influence, classifying the original image through a convolutional neural network to obtain the wearing color, style, fabric and style of the customer, and matching the clothing example with the customer
The marketplace awareness module 101 also includes a customer behavior detection module. The customer behavior detection module is used for judging whether a customer enters a fitting room or not through a pedestrian re-identification technology and recording the times of the customer entering the fitting room. The label on the clothes label of the fitting room is identified by using the radio frequency technology, and the fitting list taken by the customer is obtained. The fitting list includes the identified number of garments and the garment attributes.
The mall operation big data module 102 is used for counting based on systems such as mall video analysis and cashier desk system analysis through long-term real-time data counting and accumulated big data. And acquiring operation data of the shopping mall, wherein the operation data comprises sales volume data of shops inside the shopping mall.
The data calculation and analysis module 103 projects and transforms the customer information to a preset plan overhead image of the shopping mall, and obtains a heat map of the shopping mall area through the projection point.
Preferably, the data calculation and analysis module 103 includes a heat map acquisition module. The heat map acquisition module is used for carrying out four-point method estimation on the ground label of the image acquired by the monitoring equipment and the corresponding coordinate of the overlook image of the market plane, and carrying out projection transformation on the center point of the bottom edge of the target bounding box through the homography matrix. And generating a two-dimensional Gaussian distribution thermodynamic diagram based on projection points in the planar top-view image of the mall. The gaussian kernel size is set to 3 x 3 in an embodiment of the invention. And adding the thermodynamic diagrams between the frames according to pixels, obtaining a projection heat map of the market area through time sequence statistics, wherein the pixel value of each pixel point in the projection heat map is the heat of the position. In the embodiment of the invention, the cycle time set by the time sequence statistics is one hour.
Preferably, the data calculation and analysis module 103 further comprises an information filtering module. And the information filtering module is used for filtering the projection heat map by adopting a grid point sampling method. And processing the projection heat map through a set sliding window, wherein the sliding window only keeps the maximum value in the window every time, and other values are subjected to zero returning processing. The information filtering module can filter most data and reduce subsequent calculation amount. In the embodiment of the present invention, the window size of the sliding window is 5 × 5, and the step size is 5.
The data calculation and analysis module 103 creates a Thiessen polygon by projecting discrete data points representing heat in the heat map. The theoretical area of influence for each discrete data point can be determined by constructing a Thiessen polygon. The steps of the Thiessen polygon are as follows:
1) and automatically constructing a triangulation (Delaunay) triangulation network by the discrete data points. The discrete data points and the formed triangles are numbered and it is recorded which three discrete data points each triangle is made up of.
2) The numbers of all triangles adjacent to each discrete data point are found and recorded. I.e. find all triangles with one and the same vertex in the constructed triangulation.
3) And sorting the triangles adjacent to each discrete data point in a clockwise or counterclockwise direction for further connection to generate a Thiessen polygon. Let the discrete data point be o. Finding out a triangle with o as a vertex and setting the triangle as A; taking another vertex of the triangle A except o as a, and finding out another vertex as f; the next triangle must be bounded by of, which is triangle F; the other vertex of the triangle F is e, and the next triangle takes oe as the side; this is repeated until the oa edge is reached.
4) And calculating the center of a circumscribed circle of each triangle and recording.
5) And connecting the centers of the circumscribed circles of the adjacent triangles according to the adjacent triangles of each discrete data point to obtain the Thiessen polygon. For the Thiessen polygon at the edge of the triangular net, a vertical bisector can be made to intersect with the figure outline to form the Thiessen polygon together with the figure outline.
The data calculation and analysis module 103 calculates the vertex heat value of each vertex in the Thiessen polygon through a spatial interpolation algorithm. The method specifically comprises the following steps:
a) the vertex heat value of the common vertex of at least two Thiessen polygons in the Thiessen polygon vertices is preferably calculated. The distance of each Thiessen polygon vertex (x1, y1) to a discrete data point (x, y) within an adjacent Thiessen polygon is calculated. Calculating a distance weight for each vertex, the weight being expressed in terms of the inverse of the distance:
Figure BDA0002888272170000071
wherein the content of the first and second substances,
Figure BDA0002888272170000072
denotes the distance from the ith vertex to the discrete data point, n denotes the total number of vertices of the Thiessen polygon where the discrete data point is located, λiA distance weight for each vertex.
Calculating the final Thiessen polygon vertex heat degree through the obtained weights:
Figure BDA0002888272170000073
wherein the content of the first and second substances,
Figure BDA0002888272170000074
the heat value of the vertex of the Thiessen polygon is obtained. F (X)i,Yi) Representing discrete data point values within the ith adjacent Thiessen polygon.
b) When the vertex of the Thiessen polygon only belongs to one Thiessen polygon, namely only one discrete data point influences the vertex, the vertex is positioned at the boundary of the image, and the method for obtaining the heat degree of the vertex comprises the following steps:
firstly, obtaining the average heat attenuation proportion of the polygon:
Figure BDA0002888272170000075
wherein k is the number of vertexes of the polygon after spatial interpolation,
Figure BDA0002888272170000076
for the ith vertex heat value after spatial interpolation, dlThe distance from the ith vertex which has been spatially interpolated to the discrete data point of the polygon. Bx is the average heat decay fraction of the ith Thiessen polygon. HiI.e., the discrete data point heat value of the ith Thiessen polygon.
Then, the heat value of the vertex which is not interpolated in the polygon is calculated:
Figure BDA0002888272170000077
wherein d isOThe distance from the vertex before spatial interpolation in the No. O polygon to the discrete data point of the polygon.
Figure BDA0002888272170000078
The heat value of the un-interpolated vertex in the No. O polygon.
After the data calculation and analysis module 103 obtains the heat value of each vertex of the Thiessen polygon, the distance from any point in the Thiessen polygon to the vertex of the Thiessen polygon is calculated. And (c) calculating the heat of the point by using the weight calculation formula and the heat calculation formula in the step a).
The influence range of the heat point is not considered by directly interpolating the heat value of each point through the IDW, the vertex of the Thiessen polygon is commonly owned by a plurality of Thiessen polygons, and the direct interpolation has errors and further influences the precision. And distributing the heat value of the pixels in the area to each pixel according to the proportion of the reciprocal (or the square of the reciprocal) of the distance without considering the heat value of the adjacent area, so that the heat value of the points which are farther away from the discrete data points in the polygonal area is lower, and the rules of heat statistics are not met. Meanwhile, the heat value is reflected by Gaussian distribution of the human projection position and time sequence superposition statistics, and discrete data points sampled by the maximum grid point may cause large heat value difference values among the formed Thiessen polygons. Therefore, the embodiment of the invention adopts double IDW to interpolate the heat value of the vertex of the Thiessen polygon, and then interpolates the heat value of any position in each Thiessen polygon area through the vertex heat value. The reliability and the accuracy of the calculation result are ensured.
Preferably, the data calculation and analysis module 103 further comprises a region heat acquisition module. The area heat acquisition module is used for carrying out classification area division on the top plan view of the shopping mall to obtain an area image. Calculating the area of any region in the region image, and calculating the region heat through the number and the area of Thiessen polygons contained in the region and a point value of discrete data points:
Figure BDA0002888272170000081
wherein, AreaHiIndicating the heat, S, of the ith area in the area imagei,jDenotes the overlapping area of the jth Thiessen polygon contained in the ith area, SiDenotes the area of the i-th region in the region image, Hi,jAnd the heat value of the discrete data points of the jth Thiessen polygon contained in the ith area is represented, and n represents that n areas are divided in the area image.
The data calculation and analysis module 103 further comprises a cluster analysis module. And the clustering analysis module is used for generating a market heat map after obtaining the heat of any point. And processing the pixels with the pixel values larger than 0 in the market heat map by using a Kmeans clustering algorithm. In the embodiment of the present invention, the clustering center number of Kmeans is 3, which represents that low heat, medium heat, and high heat are classified based on the pixel value, and the corresponding heat levels are represented by numbers 1, 2, and 3, respectively. And generating a heat level characteristic diagram.
The data calculation and analysis module 103 constructs the sales data obtained by the shopping mall operation big data module 102 into a sales sequence of the clothing region. And acquiring a sales volume proportion sequence of each clothing region by using a Softmax function. And forming a clothing region sales volume proportion matrix, wherein the matrix has the same size with the heat level characteristic diagram. And combining the clothing region sales ratio matrix and the heat level characteristic diagram to obtain a shop region characteristic matrix.
The goods pushing module 104 is configured to combine the customer information and the store area feature matrix for analysis, and push the goods to the customer through the display device.
Preferably, the goods pushing module 104 includes a goods pushing neural network module. Referring to fig. 2, a block diagram of a commodity push neural network according to an embodiment of the present invention is shown. And the commodity pushing neural network module is used for analyzing the client physical information client behavior information, the client clothing matching information and the shop area characteristic matrix which are input into the network through a pre-trained commodity pushing neural network and outputting the pushed commodity.
The commodity push neural network training process specifically comprises the following steps:
1) the data is first integrated.
And constructing a one-dimensional matrix 201 of the physique of the customer, and obtaining the sex, age, weight grade and height of the customer according to the physique information of the customer. And forming the client constitution information into a one-dimensional matrix, wherein the shape of the matrix is [1,4], 4 respectively represents the client constitution information, and 1 represents a matrix.
And (3) constructing a customer behavior one-dimensional matrix 203, acquiring the times of trying on and off the fitting room, the times of changing clothes, the attributes of changing clothes and the position of the customer through the customer behavior information, and generating a one-dimensional matrix, wherein the shape of the matrix is [1,4], 4 respectively represents the customer behavior information, and 1 represents one matrix. The customer location (X, Y coordinates) refers to the customer projected coordinates in the store's plan top view image.
Constructing a customer clothing matching matrix 202, obtaining the color, style, fabric and style of the customer clothing through the customer clothing matching information, and generating a one-dimensional matrix, wherein the shape of the matrix is [1,4,5], 4 represents four descriptions of the clothing matching information, 5 represents a clothing type index value, and 1 represents a matrix.
In the embodiment of the present invention, the structure of the matrix is illustrated by a customer clothing matching matrix, the colors in the customer clothing matching matrix 202 represent the maximum color proportion of the clothing, including red (1), yellow (2), black (3), white (4), blue (5), brown (6), etc., and the numbers in the brackets represent the index values of different colors; the styles comprise occupation (1), fashion (2), elegance (3), printing (4), leisure (5), evening dress (6) and the like, and numbers in brackets represent index values of the same style; the fabric represents the maximum fabric proportion of the garment and comprises cotton (1), hemp (2), silk (3), chemical fiber (4) and the like, and numbers in brackets represent index values of different fabrics; the styles include sweet (1), japanese (2), europe and america (3), english (4), etc., and the numbers in parentheses represent index values of different styles, and the index value is 0 when the customer does not wear the hat scarf. The customer clothing match matrix 202 is shown in table 1, and the values of the matrix are index values:
TABLE 1 clothing collocation matrix
Figure BDA0002888272170000091
Figure BDA0002888272170000101
2) When the customer constitution one-dimensional matrix 201, the customer behavior one-dimensional matrix 203 and the customer clothing matching matrix 202 are input into the network, normalization processing is performed, and network convergence is accelerated.
The customer physique one-dimensional matrix 201 inputs the shape of the first fully-connected network 205 as [ B,4], B is the batch size of the network input, 4 is 4 attributes of the customer, and finally the first fully-connected network 205 outputs a 64-dimensional high-dimensional feature vector to the first embedding layer 209.
The customer behavior one-dimensional matrix 203 is input into the first fully-connected network 205 with the shape of [ C,4], C is the batch size of the network input, 4 is four features brought by the customer behavior, and finally the second fully-connected network 207 outputs a 64-dimensional high-dimensional feature vector to the second embedding layer 210.
The customer clothing matching matrix 202 is input into a one-dimensional convolutional neural network 206 to perform feature extraction and down sampling to obtain a feature map, and then the feature map is input into a full connection layer through a flattening (Flatten) operation and is mapped into a 64-dimensional high-dimensional feature vector to be mapped into a third embedding layer 211.
The shop area feature matrix 204 is input into a two-dimensional convolutional neural network 208 for feature extraction and downsampling to obtain a feature map, and then the feature map is input into a full-connected layer through a flattening (Flatten) operation and is mapped into a 64-dimensional high-dimensional feature vector to be mapped into a fourth embedding layer 212.
The first fully connected network 205 and the second fully connected network 207 perform a mapping function, the final number of neurons is N, with N representing the output dimension. Each fully-connected network design should be two or more layers to ensure that the sequence tensor can be adequately mapped to the eigenspace. In the present example N is 64.
The first embedding layer 209, the second embedding layer 210, the third embedding layer 211, and the fourth embedding layer 212 perform a dot product operation, and a 64-dimensional vector is input into the fifth embedding layer 213. The fifth embedding layer 213 integrates behavior information, attribute information, clothing collocation of the customer and heat information of each area inside the shop, so that similarity measurement can be better performed on the customer, and clothing recommendation can be more accurately realized.
3) The network training label data are information data obtained by customers at each time of purchase and are obtained based on online video perception, intelligent Internet of things perception and cashier statistics. The information data at each customer purchase is a category. The information data of each time of customer purchase comprises customer attributes, clothing collocation and behavior information data.
4) The network training method uses the AM-softmax loss function to carry out classification training, removes the last classification layer of the third fully-connected network 214 from the trained network, and selects the last hidden layer to output as the characteristic of data. Two data features are calculated using cosine similarity. The classification layer outputs the probability of each class, and the classification layer finally adopts a Softmax classification function.
Preferably, the number of neurons in the last classification layer of the third fully-connected network 214 is modified and trained based on the previous network, so that the network update can be realized based on the update of the commodity, thereby realizing the dynamic change of the neural network based on the clothing, further improving the accuracy of the network and realizing the function of online learning.
The working process of the commodity pushing neural network specifically comprises the following steps: after a customer enters a shop, the second fully-connected network 207 is closed, the third embedding layer 211 is set to be all 1, finally, input information of the first embedding layer 209 and the second embedding layer 210 is obtained through the first fully-connected network 205 and the one-dimensional convolutional neural network 206, and finally, after the click-taking is carried out, the input information is input into the third fully-connected network 214 to carry out similarity evaluation on purchase information data of historical customers, and the first garment preference recommendation is carried out. The data information of the first embedding layer 209 and the second embedding layer 210 is stored, then the first fully-connected network 205 and the one-dimensional convolutional neural network 206 are closed, real-time reasoning of the second fully-connected network 207 is carried out by sensing behavior characteristics of a customer through monitoring equipment, an intelligent Internet of things and other equipment, the data of the third embedding layer 211 which changes dynamically is obtained to carry out similarity evaluation on purchase information data of historical customers, and follow-up dynamic clothing recommendation based on behaviors of the customer is carried out. For the fourth embedding layer 212, since the internal heat map of the store is time-series statistical, it is sufficient to reason once every statistical period, and then save the fourth embedding layer 212 and close the two-dimensional convolutional neural network 208.
The goods pushing module 104 performs similarity evaluation on the customer information data through the goods pushing module, and then may recommend the historical clothing purchased by the similar customer information data of Top-K to the customer. In the present example K was taken to be 3. Meanwhile, the pushed commodity result is displayed through an LED screen or other display equipment, and the characteristics of the commodity can be explained through broadcasting through voice equipment, so that customers can be guided to purchase the commodity.
Preferably, the goods pushing module 104 further comprises an offline goods pushing module. The off-line commodity pushing module is used for counting the heat information H of the customer in each clothing area by analyzing the movement track information of the customer in the market after leaving the marketi：
Hi＝ti*AreaHi
Wherein HiIndicating the heat, t, of the customer in the ith clothing regioniIndicates the residence time, area H, of the customer in the ith clothing regioniThe heat value of the ith clothing region.
Then, the commodities of the region with the heat Top-K of the customer in the clothing region of the shop are selected for pushing. The customer's heat at various locations within the clothing area of the Top-K hot store may be selected for further analysis for more accurate merchandise push.
The offline commodity pushing module is used for installing APP capable of identifying Bluetooth beacons of stores on the mobile phone and pushing commodities on the mobile phone.
In summary, in the embodiment of the invention, the customer characteristic information is constructed through the visual perception inside the market and the big data perceived by the intelligent internet of things, the real-time commodity pushing is performed on the customer through the commodity pushing network in the commodity pushing module, and after the customer leaves the market, the off-line pushing module can also perform the commodity pushing on the customer on the mobile phone by analyzing the moving track and the staying time of the customer in the market, so that the purchasing power of the customer is improved.
It should be noted that: the precedence order of the above embodiments of the present invention is only for description, and does not represent the merits of the embodiments. And specific embodiments thereof have been described above. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims may be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing may also be possible or may be advantageous.
The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.

Claims (9)

1. The utility model provides an accurate screening push system of marketing based on big data and intelligent thing networking which characterized in that, the system includes: the system comprises a market perception module, a market operation big data module, a data calculation analysis module and a commodity pushing module;
the shopping mall sensing module is used for acquiring customer information through monitoring equipment in a shopping mall; the customer information includes customer location information;
the mall operation big data module is used for acquiring operation data of the mall, and the operation data comprises sales volume data of shops inside the mall;
the data calculation and analysis module is used for projecting and transforming the customer position information to a preset shopping mall plane overlook image and obtaining a projection heat map of the shopping mall area through a projection point; discrete data points representing heat are included in the projected heat map; building a Thiessen polygon through the discrete data points; calculating the peak heat value of each peak in the Tassen polygon through an interpolation algorithm; obtaining the influence degree of the vertex on the heat of any position by calculating the distance of the vertex heat value to any position in the heat map so as to obtain the heat of any position; obtaining a heat grade characteristic diagram of the shopping mall according to the heat of the any position; converting the sales data into a sales proportion matrix; combining the sales rate proportion matrix and the heat level characteristic map into a shop area characteristic matrix;
the commodity pushing module is used for pushing commodities to the customer through display equipment in combination with the customer information and the shop area characteristic matrix analysis.
2. The marketing accurate screening and pushing system based on big data and intelligent internet of things according to claim 1, wherein the market perception module further comprises a customer behavior detection module;
the customer behavior detection module is used for judging whether the customer enters a fitting room or not through a pedestrian re-identification technology and recording the times of the customer entering the fitting room; identifying the label brought into the clothing hangtag of the fitting room by using a wireless video technology to obtain a fitting list taken by the customer; the fitting list comprises the identified number of clothes and the clothes attribute; the times of the customer entering the fitting room and the fitting list are used as customer behavior information; the customer information includes the customer behavior information.
3. The marketing accurate screening and pushing system based on big data and intelligent internet of things according to claim 1, wherein the market perception module further comprises a customer enclosure acquisition module and an enclosure analysis module;
the customer enclosure frame acquisition module is used for outputting a target enclosure frame of the customer through a pre-trained target detection network;
the bounding box analysis module is used for analyzing the target bounding box through a pre-trained convolutional neural network to obtain the physique information of the customer; the customer information comprises the customer physique information; obtaining customer clothing matching information through a pre-trained example segmentation network; the customer information comprises the customer clothing matching information.
4. The marketing precise screening and pushing system based on big data and intelligent internet of things according to claim 1 or 3, wherein the data calculation and analysis module further comprises a heat map acquisition module;
the heat map acquisition module is used for carrying out four-point method estimation on the ground labeling of the image acquired by the monitoring equipment and the corresponding coordinates of the top plan image of the shopping mall, projecting the bottom edge center of the target enclosure frame into the top plan image of the shopping mall through a homography matrix and acquiring the projection point; generating a thermodynamic diagram with two-dimensional Gaussian distribution based on the projection points in the planar top-view image of the mall; and obtaining the projection heat map of the market area according to time sequence statistics.
5. The marketing accurate screening and pushing system based on big data and intelligent internet of things according to claim 1, wherein the data calculation and analysis module further comprises an information filtering module;
the information filtering module is used for filtering the heat map by a maximum lattice point sampling method; and processing the projection heat map through a preset sliding window, wherein only the maximum value in the sliding window is reserved in each processing, and other values are subjected to zero returning processing.
6. The marketing accurate screening and pushing system based on big data and intelligent internet of things according to claim 1, wherein the data calculation and analysis module further comprises a regional heat acquisition module;
the region heat acquisition module is used for carrying out classification region division on the top plan view of the market to obtain a region image; calculating the area of any region in the region image, and calculating the region heat through the number, the area and discrete data point values of the Thiessen polygons contained in the region:
Figure FDA0002888272160000021
wherein, AreaHiRepresenting the heat, S, of the ith area in the area imagei,jRepresenting the area of the jth Thiessen polygon in the ith area, SiRepresenting the area of the ith said region in said region image, Hi,jDiscrete data point values representing the jth Thiessen polygon contained in the ith region, n representing n regions divided in the region image.
7. The marketing accurate screening and pushing system based on big data and intelligent internet of things according to claim 1, wherein the data calculation and analysis module further comprises a cluster analysis module;
the cluster analysis module is used for generating a market heat map after obtaining the heat of any position, and dividing the heat level of the market heat map based on pixel values through a clustering algorithm to generate the heat level characteristic map.
8. The marketing accurate screening and pushing system based on big data and intelligent internet of things according to claim 1, wherein the commodity pushing module further comprises a commodity pushing neural network module;
and the commodity pushing neural network module is used for analyzing the customer information and the shop area characteristic matrix input into the network through the trained commodity pushing neural network and outputting the pushed commodities.
9. The marketing accurate screening and pushing system based on big data and intelligent internet of things according to claim 1, wherein the commodity pushing module further comprises an offline commodity pushing module;
the off-line commodity pushing module is used for counting the heat and the stay time of the area in the market where the track information is located by analyzing the track information of the customer in the market, and pushing the commodities through a mobile phone after the customer leaves the market.
CN202110019917.6A 2021-01-07 2021-01-07 Marketing accurate screening push system based on big data and intelligent internet of things Active CN112598486B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110019917.6A CN112598486B (en) 2021-01-07 2021-01-07 Marketing accurate screening push system based on big data and intelligent internet of things

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110019917.6A CN112598486B (en) 2021-01-07 2021-01-07 Marketing accurate screening push system based on big data and intelligent internet of things

Publications (2)

Publication Number Publication Date
CN112598486A true CN112598486A (en) 2021-04-02
CN112598486B CN112598486B (en) 2023-08-11

Family

ID=75207887

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110019917.6A Active CN112598486B (en) 2021-01-07 2021-01-07 Marketing accurate screening push system based on big data and intelligent internet of things

Country Status (1)

Country Link
CN (1) CN112598486B (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113344603A (en) * 2021-04-16 2021-09-03 华蔚集团(广东)有限公司 Intelligent marketing system for clients served after class
CN115564472A (en) * 2022-09-22 2023-01-03 内蒙古宏联网络科技有限责任公司 Equipment interactive information service system and method based on big data
WO2023108993A1 (en) * 2021-12-15 2023-06-22 平安科技（深圳）有限公司 Product recommendation method, apparatus and device based on deep clustering algorithm, and medium
CN116629982A (en) * 2023-07-24 2023-08-22 北京一凌宸飞科技有限公司 Network marketing method based on big data analysis
CN116958908A (en) * 2023-09-19 2023-10-27 坤前计算机有限公司 Monitoring data processing method and system

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20150339952A1 (en) * 2014-05-24 2015-11-26 Nirit Glazer Method and system for using location services to teach concepts
WO2016025321A1 (en) * 2014-08-13 2016-02-18 OneCloud Labs, Inc. Replication of virtualized infrastructure within distributed computing environments
WO2017054041A1 (en) * 2015-09-28 2017-04-06 Beehaviour.Net Pty Ltd Method, system and computer program for recording online browsing behaviour
CN109189949A (en) * 2018-08-03 2019-01-11 江苏省城市规划设计研究院 A kind of population distribution calculation method
CN111988734A (en) * 2020-08-28 2020-11-24 北京思特奇信息技术股份有限公司 Regional marketing method and system based on thermodynamic diagram
CN113344603A (en) * 2021-04-16 2021-09-03 华蔚集团(广东)有限公司 Intelligent marketing system for clients served after class
CN113506144A (en) * 2021-09-03 2021-10-15 海门市创睿机械有限公司 Clothing sales prediction method and system based on artificial intelligence and big data

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20150339952A1 (en) * 2014-05-24 2015-11-26 Nirit Glazer Method and system for using location services to teach concepts
WO2016025321A1 (en) * 2014-08-13 2016-02-18 OneCloud Labs, Inc. Replication of virtualized infrastructure within distributed computing environments
WO2017054041A1 (en) * 2015-09-28 2017-04-06 Beehaviour.Net Pty Ltd Method, system and computer program for recording online browsing behaviour
CN109189949A (en) * 2018-08-03 2019-01-11 江苏省城市规划设计研究院 A kind of population distribution calculation method
CN111988734A (en) * 2020-08-28 2020-11-24 北京思特奇信息技术股份有限公司 Regional marketing method and system based on thermodynamic diagram
CN113344603A (en) * 2021-04-16 2021-09-03 华蔚集团(广东)有限公司 Intelligent marketing system for clients served after class
CN113506144A (en) * 2021-09-03 2021-10-15 海门市创睿机械有限公司 Clothing sales prediction method and system based on artificial intelligence and big data

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
贺东: "卷烟营销网建动态监管服务体系研究", 《全国流通经济》, pages 16 - 17 *

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113344603A (en) * 2021-04-16 2021-09-03 华蔚集团(广东)有限公司 Intelligent marketing system for clients served after class
WO2023108993A1 (en) * 2021-12-15 2023-06-22 平安科技（深圳）有限公司 Product recommendation method, apparatus and device based on deep clustering algorithm, and medium
CN115564472A (en) * 2022-09-22 2023-01-03 内蒙古宏联网络科技有限责任公司 Equipment interactive information service system and method based on big data
CN116629982A (en) * 2023-07-24 2023-08-22 北京一凌宸飞科技有限公司 Network marketing method based on big data analysis
CN116629982B (en) * 2023-07-24 2023-09-29 北京一凌宸飞科技有限公司 Network marketing method based on big data analysis
CN116958908A (en) * 2023-09-19 2023-10-27 坤前计算机有限公司 Monitoring data processing method and system
CN116958908B (en) * 2023-09-19 2023-12-15 坤前计算机有限公司 Monitoring data processing method and system

Also Published As

Publication number Publication date
CN112598486B (en) 2023-08-11

Similar Documents

Publication Publication Date Title
CN112598486B (en) Marketing accurate screening push system based on big data and intelligent internet of things
US11574187B2 (en) Pedestrian attribute identification and positioning method and convolutional neural network system
CN108549891B (en) Multi-scale diffusion well-marked target detection method based on background Yu target priori
TWI559242B (en) Visual clothing retrieval
CA2734143C (en) Method and apparatus for estimating body shape
CN108062525B (en) Deep learning hand detection method based on hand region prediction
Zhang et al. A novel graph-based optimization framework for salient object detection
CN110147743A (en) Real-time online pedestrian analysis and number system and method under a kind of complex scene
CN110334687A (en) A kind of pedestrian retrieval Enhancement Method based on pedestrian detection, attribute study and pedestrian's identification
CN111784845B (en) Virtual try-on method and device based on artificial intelligence, server and storage medium
CN106951870A (en) The notable event intelligent detecting prewarning method of monitor video that active vision notes
CN108537166A (en) The method and apparatus for determining shelf pageview and analyzing pageview
CN111738344A (en) Rapid target detection method based on multi-scale fusion
CN115240119A (en) Pedestrian small target detection method in video monitoring based on deep learning
CN112819001B (en) Complex scene cigarette packet recognition method and device based on deep learning
CN112766145B (en) Method and device for identifying dynamic facial expressions of artificial neural network
CN111310531B (en) Image classification method, device, computer equipment and storage medium
Wan et al. Mixed local channel attention for object detection
CN112232334B (en) Intelligent commodity selling identification and detection method
Hao et al. Structure-based object detection from scene point clouds
Liu et al. RGB-D action recognition using linear coding
CN113657375B (en) Bottled object text detection method based on 3D point cloud
CN105955473A (en) Computer-based static gesture image recognition interactive system
Kröckel et al. Customer tracking and tracing data as a basis for service innovations at the point of sale
Saraee et al. Savoias: A diverse, multi-category visual complexity dataset

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant